Procedure for verification of sound source coverage over facades according to the International Standard ISO 140-5 Jose Luis Sanchez Bote , Antonio Pedrero Gonzalez, Juan Jose Gomez Alfageme DIAC-E.U.I.T. Telecomunicacion, Universidad Politecnica de Madrid, Ctra. Valencia, km 7, 28031 Madrid, Spain

A B S T R A C T

This paper presents a new verification procedure for sound source coverage according to ISO 140-5 requirements. The ISO 140-5 standard applies to the measurement of facade insulation and requires a sound source able to achieve a sufficiently uniform sound field in free field conditions on the facade under study. The proposed method involves the electroacoustic characterisation of the sound source in labora­tory free field conditions (anechoic room) and the subsequent prediction by computer simulation of the sound free field radiated on a rectangular surface equal in size to the facade being measured. The loud­speaker is characterised in an anechoic room under laboratory controlled conditions, carefully measuring directivity, and then a computer model is designed to calculate the acoustic free field coverage for differ­ent loudspeaker positions and facade sizes. For each sound source position, the method provides the maximum direct acoustic level differences on a facade specimen and therefore determines whether the loudspeaker verifies the maximum allowed level difference of 5 dB (or 10 dB for facade dimensions greater than 5 m) required by the ISO standard. Additionally, the maximum horizontal dimension of the facade meeting the standard is calculated and provided for each sound source position, both with the 5 dB and 10 dB criteria. In the last section of the paper, the proposed procedure is compared with another method used by the authors in the past to achieve the same purpose: in situ outdoor measure­ments attempting to recreate free field conditions. From this comparison, it is concluded that the pro­posed method is able to reproduce the actual measurements with high accuracy, for example, the ground reflection effect, at least at low frequencies, which is difficult to avoid in the outdoor measure­ment method, and it is fully eliminated with the proposed method to achieve the free field requisite.

1. Introduction

Section 4.2 of the International Standard ISO 140-5, "Field mea­surements of airborne sound insulation of facade elements and facades" [1-3], specifies that the directivity of the loudspeaker used in a test must ensure local level differences less than 5 dB (or 10 dB for facade dimensions greater than 5 m), as measured in a free field over an area the same size and orientation as the wall or element to be tested. This requirement must be verified in the frequency bands of interest defined as, at a minimum, the third oc­tave bands from 100 Hz to 3150 Hz, and preferably from 50 Hz to 5 kHz. In addition, the ISO/IEC 17025 standard, "General require­ments for the competence of testing and calibration laboratories", in Section 5.5.2 about equipment [4], requires accredited laborato­ries to validate their equipment compliance with all technical requirements of the testing standards, and compliance testing

should be performed periodically if the equipment is susceptible to changing its characteristics over time.

As stated in the ISO 140-5 standard, the uniformity level condi­tion of the facade must be measured in free field conditions, but the verification is difficult to perform in actual conditions because reflections from nearby surfaces are difficult to avoid. In recent years, the authors have performed multiple insulation measure­ments of building facades according to the ISO 140-5 standard and, before development of the procedure described in this paper, the verification of the sound source coverage on the specimen sur­face under study has always been a prohibitively expensive task, providing neither robust nor repeatable results. In those cases, the coverage specification was measured on an imaginary surface having the same dimensions as the facade being tested, and situ­ated in an outside location, with ground sound reflection always present and too great to produce a sound free field on the surface under consideration.

Accounting for the abovementioned difficulties, the authors have recently implemented a method at the Laboratory of Acous­tics of the EUIT Telecomunicacion (Universidad Politecnica de Madrid) in which the direct sound field on a surface, in the layout

indicated in the ISO standard, is calculated from directivity mea­surements of loudspeakers performed in an anechoic room [5] and used to verify the sound source.

Some other strategies exist to verify the coverage of a sound source on a facade in an anechoic room. For example, verification can occur through the actual geometry, which would require a large (and usually infeasible) anechoic room, or through scaled-down geometry, where small inaccuracies in the microphone posi­tions can generate large errors in the results. There is no flexibility in these two procedures to extend the findings to other surface sizes or loudspeaker positions.

The verification procedure developed in this study has two parts. The first part includes the measurement and characterisation of the sound source. In the second part, calculations and predic­tions are made for the specified facade. Section 2 describes the measurement stage, performed over the loudspeaker in free field conditions (i.e., an anechoic room) to determine sound directivity in the horizontal and vertical planes. The frequency response of the source is also measured. Section 3 focuses on the prediction step that consists of estimating the direct sound field on an imag­inary surface identical to that described in ISO 140-5 and with var­iable dimensions. If no specific facade size is required, the sound level estimation will be performed on the size considered standard in the publication, a surface having the horizontal and vertical dimensions Ax x Ay = 4 m x 3 m. In addition, the procedure in­cludes the estimation of the maximum horizontal dimension Axmax

of the rectangular facade that, with a variable aspect ratio of Ax/ Ay = 1, 4/3, 3/2 and 2, verifies either the 5 dB or 10 dB criterion of the ISO standard relative to the maximum direct sound level dif­ferences allowed on the facade. Finally, Section 4 compares the presented coverage verification procedure, based on computer pre­diction, with the older method, based on in situ outdoors measure­ments in acoustic free field conditions, used by the authors in the past.

2. Measurements of the sound source in an anechoic room

2.1. Measurement scenario

The measurements over the sound source are performed in the anechoic room of the Laboratory of Acoustics of the EUIT Telecom-unicacion (Universidad Politecnica de Madrid). Pictures of the measurement scenario corresponding to a test case are shown in Fig. 1. The sound source is placed on a diagonal of the anechoic room, 1.5 m from the absorbent wedges of one of the corners of the room and with its acoustic centre at height h = 1.25 m and h = 1 m above the metallic grid floor to measure horizontal and

vertical directivity, respectively (see the arrangement scheme in Fig. 2). Although the electroacoustic characteristics of the sound source should not depend on distance, measurements are made at 1 m and 1.5 m to check the measurement's repeatability and the anechoic properties of the room.

2.2. Procedure for frequency response and horizontal-vertical directivity measurement in the 1/3 octave bands from 100 Hz to 5 kHz

Among the different methods available to perform the neces­sary measurements, the method based on pink noise excitation of the sound source was chosen due to equipment availability, although any of the usual methods from the literature are applica­ble. The measurement recording and initial data processing is per­formed with the B&K PULSE 3560 spectrum analyser.

The sound source is excited via a power amplifier with a pink noise signal, and its acoustic output is picked up by the measure­ment microphone connected to the spectrum analyser. The acous­tic spectrum is measured in one-third octave bands in different rotation positions of the speaker, covering a 360° circle on both the horizontal and vertical planes, with an accuracy of 2.5°. The flat pink noise excitation spectrum (electrical signal) is also recorded simultaneously at identical points. From these measurements, the electroacoustic transfer function SPL/2.83 V [dB/W] (conven­tional voltage for speakers, representing the electric power 1 W/ 8 Q) is calculated for each point in relation to a 1-m distance, inde­pendently of the real distance measured (Fig. 3). Using the spectra measured at each point, the horizontal and vertical directivity functions, DH (0) [dB] and Dv(cp) [dB], respectively, are calculated for each frequency band (Fig. 4).

3. Calculations and predictions of the sound source being studied based on the electroacoustic measurements

3.1. Loudspeaker simulation using electroacoustic software

The data obtained following the procedure described in Section 0 characterise the sound source electroacoustically, both in direc­tivity and in frequency response, and can be used to predict the loudspeaker behaviour over the facade specimen according to the ISO 140-5 standard. The geometry of the facade insulation test prescribed by the ISO standard is described in detail in [6]. Although this geometric information, along with the loudspeaker characterisation, may be enough to perform predictions, the authors have used the EASE software tool [7-9] to implement sound source verification in the method proposed here. Any other computer tool can be used instead of EASE if the information from

Fig. 1. Test elements for the sound source under study. The arrangements shown correspond to the positions for measuring the directivity patterns in the horizontal and vertical planes.

sound source

measurement axis

measurement microphone

(a)

o-

turntable

^ r = 1 m or 1.5 m ^

grid floor

AAAAAAAAAAAAA (c)

sound source a measurement axis

measurement microphone

turntable

r = 1 m or 1.5 m ^

o

grid floor

AAAAAAAAAAAAA

vvvvvvvvv

microphone \ _ y

vyvvvv (d)

1.5 m

Fig. 2. Positions of the sound source and microphone, (a) Side view for horizontal directivity, (b) Plan view for horizontal directivity, (c) Side view for vertical directivity, (d) Plan view for vertical directivity.

I—,130 SPL [dB/W],

1 2 4 lm (eq), measured at

lm 112

100 125 160 200 250 315 400 500 630 800 Ik 1.25k 1.6k 2k 2.5k3.15k 4k 5k 6.3k 8k 10kl2.5k 16k

f [Hz]

Fig. 3. Map of electroacoustic transfer function SPL/2.83 V [dB/W] measured in the anechoic room in 1/3 octave bands. Only the data corresponding to the horizontal plane are shown.

[6] is applied. The data, in the form of one-third octave bands, have been introduced in the EASE component "EaseSpkr" to generate a new loudspeaker model (Fig. 5) of the actual sound source mea­sured in anechoic room.

The accuracy required for directivity data in EASE is 5°. A down-sampling of the data is required because the directivity measure­ments were made with accuracy of 2.5°. Measuring directivity in the horizontal and vertical planes may be sufficient for sound sources with cylindrical symmetry because the intermediate planes can be interpolated with accuracy. This procedure is nor­mally used with the method described herein because the sound sources used for measuring the insulation on facades usually have approximately cylindrical symmetry. This assumption is not appli­cable to other cases, and measuring the directivity patterns in the intermediate planes with the required accuracy of at least 5° is

mandatory. In EASE, the interpolation of the intermediate planes can be performed using the software's own method called "elliptic lobe". Fig. 6 shows an example of horizontal-vertical polar plots in EASE after data input at the 1/3 octave band of 1 kHz and directiv­ity globes after plane interpolation using the elliptic lobe.

3.2. Prediction of the direct sound field distribution on a surface according to the ISO 140-5 standard

Using the computer-aided design tool, the model of the loud­speaker under study was placed in a scenario described by the ISO 140-5 standard. A calculation surface with the standard facade dimensions of 4 m x 3 m is modelled. The loudspeaker is aimed at the centre of the facade (not mandatory in the ISO standard) in such a way that its position has an angle of 45° with the surface,

270°

24'i"

90° 270°

180° 1000 Hz

150"

5000 Hz

Fig. 4. Horizontal and vertical polar plots, DH (8) [dB] and Dv (cf) [dB]. measured in the anechoic room. The 1 kHz and 5 kHz 1/3 octave bands are shown as an example.

1 1 V

1 1 V

.

3D Peispective x • View

r

/ \ 1 ^ E ; . 1 ™ \

( > ) 1 \ [ !

) 1 \ V -View z - View

Fig. 5. Speaker model designed with the loudspeaker software. The acoustic centre of the loudspeaker is placed at the origin.

as stated in the ISO standard, and a distance of 5 m from the facade plane (Fig. 7). This is the shortest distance to the facade allowed by the ISO standard, and it produces more level differences on the sur­face. As stated in [6], for each facade size, or more precisely for each aspect ratio Ax/Ay of its surface, there are two characteristics positions of the loudspeaker to be tested: the "most favourable" position and the "least favourable" one. The "most favourable" po­sition corresponds a priori to that position in which the loud­speaker is placed just below the centre of the facade with an azimuth angle in spherical coordinates of <j>a+ = -90°. On the other hand, the "least favourable" position has the sound source located on the facade diagonal with an azimuth angle of <j>a_ = -36.9° (for the case of 4 m x 3 m dimensions). In both cases, the emitter is placed horizontally (as if standing on the ground), simulating its position in real tests.

The direct sound pressure level emitted from the previously modelled sound source is calculated over the 4 m x 3 m surface using the computer model. The loudspeaker is driven by a flat spectrum signal (pink noise type). The chosen calculation accuracy is 4 cm, corresponding to 1/8 at 1 kHz. "Isoline" type coverage

maps are obtained and represented, and the histogram of the level distribution on the surface at each frequency band is considered. The calculations take into account the sound absorption in the air due to its physical conditions, which include relative humidity of 60%, temperature of 20° C and static atmospheric pressure of 1013 hPa. A sampling of results is shown in Fig. 8.

The final goal of the verification method presented here is to determine whether the modelled loudspeaker meets the level uni­formity specifications of the ISO standard with respect to the fac­ade under study. Consequently, the maximum variation of direct sound level on the surface must be estimated. The process followed thus far gives the results shown in Fig. 9, in which the maximum level differences in the one-third octave bands from 100 Hz to 5 kHz over the 4 m x 3 m facade are visualised. The results corre­spond to two loudspeaker positions, S+(-90°) and S_(-36.9°), otherwise known as the "most favourable" and "least favourable" emission points. The final verdict of the verification process is whether the local level differences are less than 5 dB (or 10 dB for facade dimensions greater than 5 m), and therefore whether the limits of ASPL = 5 dB or ASPL = 10 dB shown in Fig. 9 are ex­ceeded in the test results.

Although the results obtained by this procedure may be suffi­cient for verification, they are limited by the validity of a sound source for a particular facade size, in this case for a size of 4 m x 3 m. It would be interesting to generalise the study for any facade size. The maximum facade size that meets the ISO standard criteria, either with ASPL = 5 dB or ASPL =10 dB, could be studied using the same computer model. This information can be extracted by studying the direct sound field produced by the modelled loud­speaker on a large surface. The surface size is then progressively re­duced until the level uniformity required by the ISO standard is met. Fig. 10 shows the maximum horizontal dimension of the fac­ade Axmax that could be used to verify ISO 140-5, with either ASPL = 5 dB or ASPL =10 dB, considering one-third octave bands up to 5 kHz and shown compared to the facade aspect ratio Ax/ Ay. To obtain this graph, the direct level distribution on a square facade measuring 22 m x 22 m has been analysed with an accu­racy of 5 cm. Four aspect ratios representative of actual building walls have been tested: Ax/Ay = 1, 4/3, 3/2 and 2. Two loudspeaker positions have been analysed for the two crucial positions, "most favourable" (azimuth angle <j>a+ = -90°) and "least favourable" (azi­muth angles <j>a_ = -45°, -36.9°, -33.7° and -26.6°, respectively for Ax/Ay = 1, 4/3, 3/2 and 2). In all cases, the loudspeaker is placed horizontally 5 m from the facade and aimed at its centre. The max­imum dimensions obtained with a reference omnidirectional loud­speaker called a "sphere" are also represented in this figure.

Fig. 6. Example of directivity representations in the loudspeaker software after data input (frequency band 1 kHz), (a) Horizontal and vertical directivity, (b) Directivity globe after plane interpolation using the software's "elliptic lobe" method.

authors using the old method with results of the verification meth­od based on computer prediction described in this paper. These comparisons are intended to prove the superiority of the sound source verification obtained with the new method compared to the previous one, in addition to its greater flexibility and lower cost.

The authors have previously collected data outdoors from vari­ous sound sources using the method of fictitious surfaces in the arrangement required by the ISO standard. The goal of this section is to find agreement between actual in situ measurements and pre­dictions made using the new method. The two major difficulties in­volved in comparing a computer scenario to reality are finding the distance from the source to the ground and the absorption coeffi­cient of this ground surface. Both factors have been obtained in this section by trial and error, though always based on realistic values. Several comparisons between the two methods have been per­formed, and all have shown very reasonably consistent results. One of these comparisons is now presented and discussed as an example. Of all the loudspeakers checked in the past and available to be characterised by the new method, the sound source B&K 4224 is widely used by the authors to measure facade acoustic insulation according to the ISO 140-5 standard. This study looked at in situ measurement data using the B&K 4224 loudspeaker posi­tioned horizontally on the ground and aimed at the centre of a fic­titious measuring surface with dimensions of 10 m x 4 m. The plan view of the arrangement's coordinate system, reproducing the ac­tual in situ verification, is visualised in Fig. 11. The loudspeaker being tested, the model B&K 4224 labelled as S (-30°) in the figure, is placed at the specified coordinates (4.3 m, -2.5 m, 5 m) and has azimuth angle of 30°. The surface size from which SPL measure­ments have been taken measures Ax xAy = 1 0 m x 4 m . The ground is considered to be 29 cm from the loudspeaker's acoustic centre.

Fig. 12 shows the maximum level differences in the one-third octave bands from 100 Hz to 5 kHz on the facade measuring 10 m x 4 m for the B&K 4224 loudspeaker arrangement shown in Fig. 11. Three different cases are compared. In Fig. 12a computer predictions following the proposed procedure with direct sound field estimation over the surface are visualised. Fig. 12b depicts computer predictions considering the acoustic reflection in the ground, with the absorption coefficient adjusted to a = 0.5. Fig. 12c shows the results from the in situ outdoor verification with real measurements.

Fig. 13 shows three maps and histograms of sound pressure le­vel distribution on the 10 m x 4 m facade in the octave band of

Table 1 shows the maximum horizontal dimensions Axmax [m] extracted from Fig. 10 for the case of the sound source previously modelled and projected over the 22 m x 22 m facade.

4. Comparison between the coverage verification procedure based on computer prediction and the method based on in situ outdoor measurements in acoustic free field conditions

The authors found many problems in the verification of loud­speaker properties when they performed measurements of the air­borne facade sound insulation of actual buildings according to the ISO 140-5 standard. Before the implementation of the procedure described in this paper, the coverage uniformity of the sound source was tested outdoors with a setting as close as possible to an acoustic free field, measuring the sound levels produced by the source over a fictitious surface whose arrangement was equal to the in situ measurement of the final insulation test. This section compares loudspeaker verifications made previously by the

S+(-90°) = (0 m, -5 m, 5 m)

Fig. 7. Plan view in the EASE coordinate system of two emitting points with the sound source at 5 m from the facade of size Ax x Ay = 4 m x 3 m. These two positions, fully described in [6], are called the "most favourable" and "least favourable" positions: S+(-90°) = (0 m, -5 m, 5 m) and S_(-36.9°) = (4m, -3 m, 5 m), respectively.

(a)

^

1.5

1

0.5

6 0

-0.5

-1

-1.5

30

20

10

0

1000Hz

-1.5 -1 -0.5 0 x [m

0.5 ]

1 1.5 2

min = • 91.6dB

max = 97.2dB min = • 91.6dB

"

104 1.5 103 102 1 101 100 0.5 99 98 97

e 0

96 95 -0.5

94 93 -1

92 91 -1.5 90 89 60

88 87 86 85

^ 40

20

84 0

5000Hz

-0.5 0 0.5 x [ m ]

min = 92dB max = 95.7dB

M 90 91 92 93 94 95 96 97 98 99 SPL[dB]

1000Hz

90 92 93 94 95 96

5000Hz

92 93

Fig. 8. Maps and histograms of direct sound pressure level distribution on the facade. The 1/3 octave bands of 1 kHz and 5 kHz are shown as an example, (a) Loudspeaker from point S+(-90°) in Fig. 7. (b) Loudspeaker from point S_(-36.9°) in Fig. 7.

1.5 100 125 160 200 250 315 400 500 630 800 Ik 1.25k 1.6k 2k 2.5k 3.15k 4k 5k

f[Hz]

Fig. 9. Maximum level differences in the third octave bands from 100 Hz to 5 kHz on the facade tested (4 m x 3 m) for the loudspeaker under study at positions S+(-90°) and S_(-36.9°) (corresponding to the "most favourable" and "least favourable" emitting points). The two level difference limits ASPL = 5 dB and ASPL = 10 dB allowed in the ISO 140-5 are depicted here.

1 kHz. Fig. 13a shows computer prediction using EASE following the proposed procedure, considering only the direct acoustic inci­dence from the loudspeaker. Fig. 13b displays the same predictions as in (a), but considering the reflection in the ground, with the

absorption coefficient adjusted to a = 0.5. Fig. 13c represents the actual measurement coming from in situ verification with outdoor SPL measurements. In all three cases, the loudspeaker S(-30°) rep­resenting model B&K 4224 was placed as in Fig. 11 at coordinates

"sphere" with ASPL = 10 dB

"sphere" with ASPL = 5dB

loudspeaker with ASPL = 10 dB

loudspeaker with ASPL = 5 dB

. tpo-

. <Ptt

- e - t p t t

Ax/Ay

Fig. 10. Maximum horizontal dimension of the facade Axmax verifying ISO 140-5, either with the ASPL = 5 dB or ASPL = 10 dB criterion, considering one-third octave bands up to 5 kHz and shown compared to the facade aspect ratio Ax/Ay. The maximum dimensions obtained with a reference omnidirectional loudspeaker called a "sphere" are represented along with the tested loudspeaker in this graph. The azimuth angle of the "most favourable" position of the loudspeaker is the same for all aspect ratios, <pa+ = -90°. The azimuth angle for the "least favourable" position of the loudspeaker varies with the facade aspect ratio, resulting in q>a_ = -45°, -36.9°, -33.7° and -26.6° for Ax/Ay = 1, 4/3, 3/2 and 2, respectively. In all cases, the loudspeaker is placed horizontally 5 m from the facade and aimed at its centre.

Ax=10m

| Y

a X

* Z ^ \ . ]<?a = -30°

0.5 m|^ earth ground S(-30°) = (4.3 m, -2.5 m, 5 m)" 0.29 m

Fig. 11. Plan view in the proposed coordinate system of the loudspeaker being tested, model B&K 4224, emitting 5 m from the fictitious facade with size Ax x Ay = 10m x 4m. The emission point is called S(-30°) (azimuth angle of 30°) and is placed at coordinates (4.3 m, -2 .5 m, 5 m). The ground is located 29 cm from the acoustic centre of the loudspeaker.

(4.3 m, -2.5 m, 5 m). The data shown in the maps have an accuracy of 1 m.

Important conclusions can be drawn from Figs. 12 and 13. First, the in situ outdoor measurements of Fig. 13c seem to suffer the ef­fects of ground reflection [10-12], as the effects are also reflected in the computer simulation of Fig. 13b. Therefore, these measure­ments are not useful for verifying the loudspeaker according to the ISO 140-5 standard because the free field condition is manda­tory to accomplish this verification. Although ground reflection in actual measurements could be mitigated at mid- to high-range fre­quencies by incorporating an absorbent material between the sound source and the floor (a procedure that was not implemented in the measurements shown in Fig. 13c), it is very difficult to ab­sorb this ground reflection at mid- to low-range frequencies (the effect of reflections between 250 Hz and 1250 Hz is present when comparing Fig. 12a with Fig. 12b). The second conclusion to be made is that computer simulations can reproduce quite accurately what happens in a simple case such as this one when considering only ground reflection. Although adjustments were made in the simulation process by trying to fix the distance from the loud­speaker acoustic centre to the floor and the absorption coefficient of this floor material at all frequencies of interest, a strong correla­tion between simulation and actual measurements was found in all loudspeaker simulations (including those not shown here), appear­ing as the standing waves shown in Fig. 13 for low- to mid-range frequencies.

It can be concluded that the proposed method is superior to the in situ outdoor measurement procedure. This method more reli­ably determines how the loudspeaker acoustic coverage behaves in free field conditions, as required by the ISO standard. The meth­od is also cheaper and more flexible because a speaker can be ver­ified only once and used later as needed to prove its effect on facades of multiple sizes and orientations in acoustic free field conditions.

Table 1 Based on Fig. 10, this table shows the maximum horizontal dimension Axmax [m] of the facade that meets ISO 140-5 for the loudspeaker under study up to the 5 kHz one-third octave band, with either the ASPL = 5 dB or ASPL = 10 dB criterion, for the aspect ratios Ax/Ay = 1, 4/3, 3/2 and 2. The maximum size Axmax is shown for the most and least favourable positions. The most favourable position has an azimuth angle of <?>„+ = -90° for all four aspect ratios. The least favourable position's azimuth angle varies with the aspect ratio, being <?>„_ = -45°, -36.9°, -33.7° and -26.6° for Ax/Ay = 1, 4/3, 3/2 and 2, respectively.

Ax/Ay= l 4/3

5dB 10 dB

3/2

5dB 10 dB

2

ASPL= 5dB 10 dB

4/3

5dB 10 dB

3/2

5dB 10 dB 5dB 10 dB

<pa-= ^ m a x =

-90° 1.35 m

-45° 1.20 m

-90° 3.65 m

-45° 3.05 m

-90° 1.85 m

-36.9° 1.60 m

-90° 4.85 m

-36.9° 4.20 m

-90° 1.95 m

-33.7° 1.70 m

-90° 5.20 m

-33.7° 4.50 m

-90° 2.7 m

-26.6° 2.4 m

-90° 5.5 m

-26.6° 5.5 m

25

20 =

15

5 -

4 100

• (a) Computer predictions, only direct field

• (b) Computer predictions, with ground effect

. (c) In situ verification with real data

125 160 200 250 315 400 500 630 800

f [Hz]

Ik 1.25k 1.6k 2k 2.5k 3.15k 4k 5k

Fig. 12. Maximum level differences in the one-third octave bands from 100 Hz to 5 kHz on the facade of size 10 m x 4 m for the loudspeaker at position S(-30°) = (4.3 m, -2 .5 m, 5 m) shown in Fig. 11. (a) Computer predictions following the proposed procedure with direct sound field estimation, (b) Computer predictions as in (a), but considering the reflection in the ground with the absorption coefficient adjusted to a = 0.5. (c) In situ verification with real measurements.

1000Hz SPL[dB] . - 8 7

707172 73 74 7576 77 78 79 80 8182 83 84 85

Fig. 13. Maps and histograms of sound pressure level distribution on a 10m x 4 m facade in the octave band of 1 kHz with the loudspeaker placed at coordinates S(—30°) = (4.3 m, -2 .5 m, 5 m), the same coordinate system as in Fig. 11. The data used to draw the maps have an accuracy of 1 m. (a) Computer prediction following the proposed procedure with direct sound field estimation, (b) Computer prediction as in (a), but considering the reflection in the ground, with absorption coefficient adjusted to a = 0.5 and the loudspeaker acoustic centre placed 29 cm from the floor, (c) In situ verification with real measurements.

5. Conclusions

This paper has presented a new verification procedure for sound source coverage based on the ISO 140-5 requirements. This proce­dure has been successfully used by the authors with the loud­speakers utilised for their current and past facade insulation measurements. The method consists of the electroacoustic charac­terisation of the sound source in laboratory free field conditions (an anechoic room) and the subsequent computer simulation pre­diction of the free sound field radiated on a facade. Coverage requirements for the source are difficult to validate outdoors be­cause of the need to achieve acoustic-free field conditions. Prior to using the proposed method, the authors had verified sound sources outdoors in as free a sound field as possible.

As demonstrated in the paper, ensuring free field condition out­doors is difficult because ground reflection on the facade surface cannot be completely avoided. In addition, outdoor measurements are usually not accurate enough because of the difficulty of mea­suring large surfaces. Each sound source position also requires the use of a costly measurement process in this scenario. In con­trast, the proposed method is more flexible and repeatable; it al­lows for high resolution predictions and the testing of multiple sound source positions and facades of varying sizes. The effect of unwanted reflections can be fully avoided, and the maximum size of the facade applicable to each sound source can be calculated.

These findings may assist in thinking about the usefulness of the ISO standard requirements for loudspeaker sound coverage on a facade. Noting that it is not clear if level differences on the fac­ade have a large influence on insulation measurements anyway [13], it might be better to specify the maximum angle of directivity

coverage of the sound source according to the maximum angle at which the facade is "seen" by the loudspeaker, a point stated in the geometric study in [6]. These suggestions should help to im­prove future versions of the ISO 140-5 standard regarding the insulation of facades.

References

[1 ] ISO 140-5:1998. Acoustics-measurement of sound insulation in buildings and of building elements. Part 5: field measurements of airborne sound insulation of facade elements and facades.

[2] Rasmussen B, Rindel JH. Sound insulation between dwellings - descriptors applied in building regulations in Europe. Appl Acoust 2010;71(3):171-80.

[3] Rasmussen B. Sound insulation between dwellings - requirements in building regulations in Europe. Appl Acoust 2010;71(4):373-85.

[4] ISO/IEC 17025:2005. General requirements for the competence of testing and calibration laboratories.

[5] Sanchez Bote JL, Gomez Alfageme, JJ. Procedimiento para la medicion y verification de la directividad y la cobertura de una fuente sobre un elemento de fachada de acuerdo a la norma ISO 140-5. In: Proc. Tecniacustica 2011. Caceres (Spain); October 26-28, 2011.11 pp. (in Spanish).

[6] Sanchez Bote JL, Pedrero Gonzalez A, Gomez Alfageme JJ. Influence of loudspeaker directivity and measurement geometry on direct acoustic levels over facades for acoustic insulation tests with the International Standard ISO 140-5. Appl Acoust 2012;73(4):440-53.

[7] EASE. Enhanced acoustic simulator for engineers, <http://ease.afmg.eu/>. [8] Feistel S, Ahnert W. Modeling of loudspeaker systems using high-resolution

data. J Audio Eng Soc 2007;55:571-97. [9] Feistel S, Ahnert W, Hughes C, Olson B. Simulating the directivity behavior of

loudspeakers with crossover filters. Paper 7254, 123rd convention, Audio Engineering Society. New York, USA; October 5-8, 2007.

[10] Hopkins C, Lam Y. Sound fields near building facades - comparison of finite and semi-infinite reflectors on a rigid ground plane. Appl Acoust 2009;70(2):300-8.

[11] Berardi U, Cirillo E, Martellotta F. Interference effects in field measurements of airborne sound insulation of building facades. Noise Control Eng J 2011;59(2):165-76.

[12] Berardi U, Cirillo E, Martellotta F. Measuring sound insulation of building facades: interference effects and reproducibility. In: Proc. inter noise 2010. Lisbon (Portugal); June 13-15, 2010. 10 pp.

[13] Pedrero Gonzalez A, Sanchez BoteJL, Ulin Nabatov V, Diaz Sanchidrian C. Influence of loudspeaker directivity on the measurement uncertainty of the acoustic testing of facades. In: Proc. inter noise 2011. Osaka (Japan); September 4-7,2011.6 pp.